Real-time monitoring of broiler flock's welfare status using camera-based technology

Broiler activity and occupation patterns are of special interest to farmers during visual inspection. However, this is time consuming and precision livestock farming (PLF) technologies can enable the monitoring of such key flock behavioural indicators in a continuous and automated way in the house. The aim is to show how the welfare status of the poultry flock can be evaluated by real-time monitoring of activity and occupation patterns. Four top view cameras were installed in a commercial broiler house for 9 complete growing cycles. The cameras recorded images continuously and they were translated into numerical values of activity and occupation indices each minute. Three welfare assessments were performed in weeks 3, 4 and 5 of each growing cycle according to the standardised Welfare Quality® assessment protocol for broiler chickens. A real-time dynamic model was developed to monitor and forecast the time evolution of these indices and the confidence intervals for normal behaviour over each growing cycle. Statistically relevant correlations (p